|Publication number||US8224489 B2|
|Application number||US 12/396,944|
|Publication date||Jul 17, 2012|
|Filing date||Mar 3, 2009|
|Priority date||Mar 3, 2008|
|Also published as||US20090222139, US20120283881, WO2009111489A2, WO2009111489A3|
|Publication number||12396944, 396944, US 8224489 B2, US 8224489B2, US-B2-8224489, US8224489 B2, US8224489B2|
|Inventors||Clifford C. Federspiel|
|Original Assignee||Federspiel, Corporation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (27), Non-Patent Citations (1), Referenced by (3), Classifications (12), Legal Events (2)|
|External Links: USPTO, USPTO Assignment, Espacenet|
This application claims the benefit of and priority to U.S. Provisional Patent Application No. 61/033,154, entitled “METHOD AND APPARATUS FOR COORDINATING THE CONTROL OF HVAC UNITS,” by Clifford C. Federspiel, filed Mar. 3, 2008, and U.S. Provisional Patent Application No. 61/149,273, entitled “METHOD AND APPARATUS FOR COORDINATING THE CONTROL OF HVAC UNITS,” by Clifford C. Federspiel, filed Feb. 2, 2009, the entire contents of which are incorporated herein by reference.
The present invention relates to controls for heating, ventilating, and air-conditioning (HVAC) systems, and more specifically to controls for systems that have multiple HVAC units conditioning a common space.
Modern buildings typically use HVAC systems to control indoor temperature, pressure, ventilation rate, and other variables. In many cases, these buildings have open-plan construction in which an open space is temperature-controlled and humidity-controlled by multiple HVAC units. In some cases, these HVAC units are referred to as unitary systems because the heating, cooling, humidification, and fans reside in a single packaged unit.
One application of unitary equipment is for the cooling and humidity control of data centers. In this application, it is common to have many HVAC units deployed throughout the data center. They are often floor-standing units, but may be wall-mounted, rack-mounted, or ceiling-mounted. The HVAC units often provide cooled air either to a raised-floor plenum, to a network of air ducts, or to the open air of the data center. The data center itself, or a large section of a large data center, typically has an open-plan construction, i.e. no permanent partitions separating the air in one part of the data center from the air in another part.
Another application of unitary equipment is for the cooling and humidity control of commercial buildings such as retail stores. In this application, it is common to have many HVAC units deployed on the rooftop of the building. The HVAC units typically provide cooled air either to a network of air ducts or directly into the open area of the store. The building itself, or large sections of it, typically has an open-plan construction, with no permanent partitions separating the air in one part of the building from the air in another part.
HVAC units for open-plan spaces such as data centers or retail stores are typically operated with decentralized, stand-alone controls. In the case of the data center HVAC units, it is common for each unit to operate in an attempt to control the temperature and humidity of the air entering the unit from the data center. For example, an HVAC unit may contain a sensor that determines the temperature and humidity of the air entering the unit. Based on the measurements of this sensor, the controls of that HVAC will alter operation of the unit in an attempt to change the temperature and humidity of the air entering the unit to align with the set points for that unit.
Since it is common for these units to be located in close proximity, this operating strategy often causes one unit to fight against another, resulting in oscillating temperature and humidity. In some cases, one unit may be heating while another unit is cooling, thus causing a waste by simultaneously heating and cooling. This problem is compounded by sensor errors from unit to unit, by the fact that the units may be configured with different set points (e.g. temperature that is to be maintained), and by the fact that humidity control is coupled with temperature control.
In a retail store or other commercial building with rooftop units and open-plan construction, the situation is similar. The units are operated with standalone controls where each unit attempts to control the temperature measured by one sensor located in the store. Since the open-plan construction allows free flow of air between sensors, the operation of one unit is coupled to the operation of another unit. Problems such as close sensor or unit location, sensor errors, and differing setpoints for the HVAC units can cause instability and unit-to-unit heating and cooling at the same time.
An alternative approach for data centers is to control each unit based on a sensor located in the data center instead of a single sensor located in the unit. U.S. patent application 2006/0206291 A1 to Bash et al. describes a method for assigning multiple sensors in an open-plan data center to a set of actuators, where there are typically fewer actuators than sensors. Each actuator is controlled (i.e. whether to heat or cool) based on one control sensor that is chosen from a set of sensors assigned to that actuator. Although Bash provides options for selecting the one control sensor, similar problems of close sensor or unit location, sensor errors, and differing setpoints can cause instability and unit-to-unit heating and cooling at the same time.
Another problem with decentralized control of HVAC units in data centers, retail stores, and other open-plan buildings is that the control is not fault tolerant. If one HVAC unit or one of its final control elements fails, or if the HVAC unit is intentionally brought off-line, the controls of the other units have no way to compensate for the failed or stopped unit.
Therefore, it is desirable to provide methods and systems that account for the coupling between units and provide mechanisms for fault tolerance.
Embodiments of the present invention provide systems and methods for controlling environmental maintenance (e.g. HVAC units) modules using sensors. Overlapping sets of sensors may be used to control respective HVAC units to account for the coupling effects of multiple HVAC units on a common controlled space. Embodiments can also advantageously ensure that the spatial variation of the controlled environmental variables is minimized, thus increasing efficiency and providing a comfortable, uniform temperature for workers and consumers.
For example, environmental measurements (e.g. temperature, humidity, pressure) are received from the sensors. Error or difference values for each sensor are determined in relation to reference values (e.g. temperature settings). To maintain desired environmental settings, changes in an operational parameter (e.g. fan speed) of each module are calculated from contributions of respective difference values. In one aspect, a transfer matrix may be used to weight the difference values and then each contribution may be summed to obtain an overall change in the parameter. This may be done for each parameter of each module.
According to one exemplary embodiment, a method of controlling an environmental maintenance module using a plurality of sensors is provided. A first set of environmental measurements are received. Each environmental measurement is from a respective one of the sensors. Differences between the environmental measurements and a plurality reference values are determined. Each difference is between a respective environmental measurement and a corresponding reference value. A change value for an operational parameter of the environmental maintenance module is calculated. The change value includes contributions derived from at least two of the differences.
According to another exemplary embodiment, a system for controlling an environmental maintenance module using a plurality of sensors is provided. An input interface receives a first set of environmental measurements. Each environmental measurement is from a respective one of the sensors. One or more processors are configured to determine differences between the environmental measurements and a plurality reference values. Each difference is between a respective environmental measurement and a corresponding reference value. The one or more processors are also configured to calculate a change value for an operational parameter of the environmental maintenance module. The change value includes contributions derived from at least two of the differences. An output interface sends a new value for the operational parameter to the environmental maintenance module.
According to another exemplary embodiment, a method of controlling a plurality of environmental maintenance modules using a plurality of sensors is provided. A first set of environmental measurements are received. Each environmental measurement is from a respective one of the sensors. Differences between the environmental measurements and a plurality reference values are determined. Each difference is between a respective environmental measurement and a corresponding reference value. A first change value is calculated for an operational parameter for a first environmental maintenance module. The first change value includes a first contribution derived from the difference associated with a first sensor. A second change value is calculated for the operational parameter for a second environmental maintenance module. The second change value includes a second contribution derived from the difference associated with the first sensor, where the first contribution is different from the second contribution.
A better understanding of the nature and advantages of the present invention may be gained with reference to the following detailed description and the accompanying drawings.
Embodiments provide systems and methods for controlling environmental maintenance (e.g. HVAC units) modules using sensors. To account for the coupling effects of multiple HVAC units on a common controlled space, the sensors may be composed of overlapping sets of sensors that are used to control respective HVAC units. Embodiments can also advantageously ensure that the spatial variation of the controlled environmental variables is minimized, thus increasing efficiency and providing a comfortable, uniform temperature for workers and consumers.
In one embodiment, HVAC units 2 are unitary equipment that provide airflow to the data center to cool servers in server racks 4. In one aspect, HVAC units 2 can cool, heat, humidify, or dehumidify air that passes through them. Environmental sensors 3 are devices that measure environmental parameters, such as temperature or humidity. Environmental sensors 3 can transmit measurements (also called readings) by any means, such as by wired or wireless communication means (e.g., Wi-Fi, Wimax, Zigby, or any other applicable protocol).
The supervisory controller 15 coordinates the operation of multiple HVAC units 2 by computing commands to inputs 12 of each HVAC unit 2. The commands are computed based on the environmental sensor readings from the sensors 3. The inputs 12 may correspond to a variety of different HVAC units 2 and/or devices or circuits within the HVAC units 2.
In one embodiment, input 1 of HVAC unit 1 may correspond to the operational parameter of one actuator (e.g. for a cooling valve), and the input 2 of HVAC unit 1 may correspond to a different actuator of the same HVAC unit 1. In another embodiment, some of the HVAC units 2 have only one input.
In another embodiment, one input could be the setpoint for the temperature of an HVAC unit 2. For example, a setpoint may be the desired temperature of the air discharged by the HVAC unit 2, or the desired temperature of the air returning to the unit. Other inputs could be the setpoint for the humidity (or the humidifier command), or a command to a variable frequency drive (VFD).
In one embodiment, each HVAC unit has the same number of inputs, each corresponding to one actuator of that HVAC unit. In another embodiment, different HVAC units may have a different number of actuators. In such an embodiment, the number of sensors may stay the same even though some of the HVAC units have less functionality than other ones.
In part, a reason the number of sensors may stay the same is because the sensors may affect each actuator as explained herein. For example, a temperature actuator (e.g. cooling valve) can affect the humidity as may happen when condensate forms on the cooling coil if the environment is cold enough. Likewise, humidity actuators (e.g. infrared humidifiers and evaporative cooling valves) affect the temperature, as may happen when infrared humidifiers raise humidity or evaporative coolers raise humidity.
Environmental sensors 3 send their readings back to supervisory controller 15, e.g., by wired or wireless communication means (such as Wi-Fi, Wimax, Zigby, or any other applicable protocol). Examples of sensors include temperature sensors, humidity sensors, and pressure sensors. The environmental sensors 3 may be positioned randomly or according to a regular pattern. The environmental sensors 3 may also be organized via clusters of sensors or individually placed.
In one embodiment, supervisory controller 15 causes environmental sensor readings F to be as close as possible to an associated set of reference values 8. In another embodiment, the supervisory controller 15 automatically compensates for failed HVAC units 2, failed final control elements, or intentionally stopped HVAC units 2.
In one embodiment, supervisory controller 15 internal stores the set of desired reference values 8 for each environmental sensor, e.g. in flash memory, cache, or other suitable memory. In other embodiments, the reference values 8 may be stored externally, e.g. in a disk drive or optical drive. In operation, supervisory controller 15 adjusts inputs of HVAC units 2 to keep the values from environmental sensors 3 close to their desired reference values 8.
In one embodiment, negation blocks 14 changes the sensor readings to have a negative value (e.g. by multiplying with −1). In this manner, to provide negative feedback, summation calculator 10 can take a sum instead of a subtraction to provide a difference between reference values 8 and the sensor readings.
Inputs to HVACs
In one embodiment, supervisory controller 15 computes commands that are provided to inputs 12 and are used directly for final control elements (e.g. actuators) in HVAC units 2. These commands sent to the inputs 12 may be provided, e.g., by wired or wireless communication means.
In another embodiment, supervisory controller 15 computes commands to the inputs 12 that are used by a local digital controller (e.g. having microprocessor-based controls) in an HVAC unit 2. In one aspect, each input to the local digital controller of a unit corresponds to an actuator of the unit. The local digital controller can then determine the final commands sent to the final control elements. For example, the local digital controller may convert a digital signal to an analog signal for the actuator, or convert a protocol of the signal to be usable by an actuator.
The local digital controller may also operate to maintain an actuator at a particular setting though a local control loop. Thus, supervisory controller 15 may command the setpoints of local control loops in the local digital controllers rather than directly commanding the final control elements.
In one embodiment, supervisory controller 15 has means of receiving status indicators 13 from the environmental sensors 3 and/or the HVAC units 2. The status indicators 13 provide information as to whether an HVAC unit 2 or a sensor 3 is presently operational. The status indicators 13 are shown separated from the HVAC units 2 and sensors 3 for illustrative purposes.
In one embodiment, the status indicators 13 for the HVAC units 2 may be obtained from local digital controllers of the HVAC units 2. These local digital controllers can be queried by supervisory controller 15 to determine if the local digital controllers or the HVAC units 2 are “on” or “off”. If a unit is “off”, then the status indicator 13 for that unit's actuators would be, e.g., zero.
In another example, the environmental sensors 3 have some well-defined and easily detected failure modes. In one aspect, one failure mode is an “unreachable”, which means that a gateway, e.g. a network interface of the supervisory controller 15, cannot communicate with the sensor. Another failure mode is an out-of-range voltage (either 0 volts or 1.5 volts), where 0 volts implies that the sensor probe has a short circuit and 1.5 volts indicates that the sensor probe has an open circuit or is missing. Any of these failures may result in a status indicator of zero for that sensor.
II. Translating Feedback Errors to Control Errors
In step 405, the supervisory controller 15 acquires values F from environmental sensors 3. The F values are readings from the environmental sensors 3. F may be viewed as a vector, with each element corresponding to a different reading. Note that sensors of different types may be packaged as a single unit, but are still different sensors.
In step 410, the supervisory controller 15 acquires status values from the status indicators 13. For example, the status values may be received by a pseudoinverse calculator 10. In one embodiment, the status indicators 13 are outputs of logical operations, as may result from digital and/or analog data being processed to determine whether or not a sensor or actuator is working.
In step 415, supervisory controller 15 computes the respective difference between the desired reference values R and the corresponding values F from environmental sensors 3. In one embodiment, a summation calculator 9 of the supervisory controller 15 is used to perform the difference calculation. Inputs to summation calculator 9 are desired reference values 8 and outputs from environmental sensors 3. In one aspect, the calculation is performed as follows:
E f =R−F (1)
where Ef is the feedback error, R is a vector of desired reference values 8, and F is a vector of the corresponding readings from environmental sensors 3. Outputs of summation calculator 9 are input to pseudoinverse calculator 10.
In step 420, a transfer matrix G is retrieved, e.g., from a local or external (e.g. networked) memory. G is a matrix of values that relate inputs to process 16 (e.g. inputs 12) to outputs of process 16 (e.g. outputs of sensors 3). In one aspect, the element in the ith row and jth column of G is the transfer function relating the jth final control element to the ith sensor.
In one embodiment, each matrix element is a function (e.g. a LaPlace transform). In another embodiment, the matrix element is a number (e.g. the value of the LaPlace transform when s=0). In one aspect, using numbers provides static decoupling.
In one embodiment, the values of G are equal to the change in the ith environmental sensor reading divided by the change in the jth final control element value. For example, the element of G at the ith row and jth column may be the ith output divided by the jth input.
In general, G is frequency dependent. In one embodiment, the values of G at or close to steady state (zero frequency), also called “quasi-steady,” are used. A method of creating G is described below in
In step 425, the transfer matrix G is optionally modified based on the status indicators. In an embodiment, G is multiplied element-by-element by the status matrix, S, which has the same dimension as G but contains information about the status of each final control element.
G s =G.*S (2)
where the “.*” operation signifies element-by-element multiplication.
In one embodiment, the status matrix, S, is an incidence matrix with values of unity for every element (i.e. every sensor that is working) of every column if the corresponding final control element is operational. If an HVAC unit 2 is “off”, then the columns of S corresponding to the final control elements of that unit are zero. If a sensor is “off”, then the row of S corresponding to that sensor are zero. One skilled in the art will recognize that the appropriate elements of G may be zeroed in any number of ways.
Accordingly, embodiments can advantageously compensate for failed or shut-down units automatically. In particular, embodiments provide a control strategy that can account for the effects of multiple HVAC units on a common controlled space when one or more HVAC units 2 have been taken off-line or has failed.
In step 430, an inverse (such as an exact inverse, pseudoinverse, or Moore-Penrose pseudoinverse) of Gs is computed. In one embodiment, this is calculated by singular value decomposition of Gs. The singular value decomposition of Gs is computed as follows:
G s =UΣV* (3)
where U is the matrix of left singular vectors, Σ is the diagonal matrix of singular values, V is the matrix of right singular vectors, and V* is the conjugate transpose of V. The pseudoinverse of Gs may be computed as follows:
G s + =U*Σ + V (4)
where U* is the conjugate transpose of U, and Σ is the transpose of Σ with every nonzero entry replaced by its reciprocal. In one embodiment, if a nonzero entry of Σ is smaller than a threshold, then that element of Σ+ is set to zero.
In another embodiment, the computing of the pseudoinverse uses QR factorization.
If values from status indicators 13 have changed, or if a condition has been reached that causes values of the transfer function matrix, G, to change, then supervisory controller 15 may re-compute the pseudoinverse of G with pseudoinverse calculator 10.
In step 435, supervisory controller 15 computes the control error Ec as the product of the pseudoinverse and the feedback error as follows:
E c =G s + E f. (5)
Ec is the calculated error (an example of a change value) in an operational parameter (such as a setpoint, digital control value, or analog command signal) for each of the final control elements of each of the HVAC units 2.
In step 440, the supervisory controller 15 computes the new inputs of each HVAC unit 2. In some embodiments, the values for Ec may be added (or subtracted as the case may be) from the old operational parameters to create new operational parameters. These new operational parameters may be taken as is and used for the inputs 12 into the HVAC units.
In other embodiments, the new operational parameters may be combined with previously calculated operational parameters in order to determine the new inputs. In one embodiment, this combination of operational parameters from previous iterations is performed by a proportional-integral-derivative (PID) object 11 operating on each value of the control error Ec. In
In one embodiment, a PID object 11 calculates three separate parameters: the proportional, the integral and derivative values. The proportional value determines the reaction to the current error, the integral determines the reaction based on the sum of past errors, and the derivative determines the reaction to the rate at which the error has been changing. The weighted sum of these three actions is used to adjust the process 16 via a final control element, such as the position of a control valve or the power supply of a heating element, or to reset a setpoint of an embedded (local) digital controller of an HVAC unit.
For example, the setpoint may be the desired return air temperature of a computer room air handling unit or the desired discharge air temperature of a computer room air handling unit. In embodiments, the digital controller may set any one or more of the value of the chilled water valve, the value of a reheat coil, a humidifier, and a variable frequency drive for a motor.
By “tuning” the three constants in the PID object algorithm, the PID object 11 can provide control action designed for specific process requirements. For example, the response of the PID object 11 can be described in terms of the responsiveness of the PID object 11 to an error, the degree to which the controller overshoots a setpoint, and the degree of oscillation of the process being controlled.
In an embodiment, an integrator anti-windup feature resets the integral term so that the output of the PID object 11 equals the maximum output (typically 100%) or the minimum output whenever the control action is “pushing” the output against the limit. In another embodiment, a deadband feature sets the error to zero when it is close to zero. This deadband feature is used to prevent control changes due to noise or other small perturbations in the error that aren't important. In yet another embodiment, a rate constraint feature prevents the output of the PID object 11 from changing faster than a user-defined rate, which may act in a similar fashion to non-linear damping. In yet another embodiment, an auxiliary output feature enables feedforward control.
More Sensors than Actuators
In embodiments used for data-center control, there will generally be more sensors 3 (outputs) than actuators (e.g. cooling valves and humidifiers). Here, the pseudoinverse Gs + is used to solve a least-squares problem. In one aspect, using the pseudoinverse Gs + results in a temperature distribution that has the smallest mean squared deviation from the reference values, R, at steady-state. Accordingly, embodiments can advantageously produce the control actions that result in feedback errors Ef that are minimized in the least-squares sense.
Fewer Sensors than Actuators
In another embodiment, there are fewer sensors 3 than actuators. Then, the transfer function matrix G has more columns than rows, and the pseudoinverse Gs + is not unique. In one aspect, computing G using the singular value decomposition is not only numerically robust, but also produces a “minimum norm” solution that will produce actuations that are all about the same magnitude rather than some at 100% and others at 0%. Such operation can provide increased efficiency as for many kinds of HVAC equipment. For example, it is more efficient to operate redundant actuation mechanisms (e.g., fans operating in parallel) at the same level.
Effect of Status Matrix
In embodiments using the status indicators 13, the status indicators 13 tell the system whether or not an actuator or sensor is working. If an actuator is not working, either because of a failure or because it has been taken offline or because of physics (e.g., cooling coil not cold enough to dehumidify), then the column of G for that actuator should be set to zero and the pseudoinverse recomputed. The controls will then move the temperature distribution to the new least squares feedback error given that the actuator can't be used to change the temperature distribution. If a sensor fails, then the corresponding row of G can be zeroed out and the pseudoinverse recomputed. This zeroing will have the effect of making the system not care about the temperature at the location of the failed sensor.
III. Translating Feedback Values to Control Values
In certain situation of G having zero or near zero singular values, problems can arise. One embodiment solves such problems by using a pseudoinverse of a reduced-order transfer function matrix (Gr).
Problem of Zero or Near Zero Singular Values of G
In one embodiment, the dimension of the control errors Ec equals the number of actuators. Each control error is input to a PID object 11. Due to redundancy or partial redundancy, particularly in combination with measurement noise during the functional test used to determine G, it is possible for this technique to produce undesirable results. For example, suppose that the transfer function matrix is the following:
Since there are two columns, this G corresponds to a transfer matrix for two actuators. Since there are three rows, this G corresponds to a transfer matrix for three sensors. Thus, for this G, each of the actuators has an equal effect on a respective sensor.
In this case one of the singular values of G is zero (G has rank of 1), and the pseudoinverse is as follows:
The problem in this case is that feedback loops, each potentially with an integrator (one corresponding to each row of G+) will be competing for control of one state variable. The problem gets considerably worse if G is not quite singular. Consider the following modified transfer function matrix:
In this case, the singular values are 0.89 and 0.023, so G has full column rank. However, the pseudoinverse is quite different:
Recall that G+ translates feedback errors Ef to control errors Ec. Suppose that the entire datacenter is hot by one degree. This means that all three feedback errors (Ef) will equal −1 (Ef[−1 −1 −1]T), and the control errors will be the product of G+ and Ef, which gives the following result:
In other words, the controls will drive the two actuators in opposite directions even though all of the temperatures are too high by the same modest amount.
In one aspect, we can partially fix this problem by setting the small singular value to zero. This will result in a “rank-corrected” transfer function matrix (Gc). The pseudoinverse of the rank-corrected transfer function matrix will be as follows:
This gets us back to a situation similar to the first example, where we will get two control errors of about the same size (an improvement), but we will still have two integrators competing for (approximately) one state variable. Embodiments provide a solution that produces an inverse that has the same number of rows as there are “significant” singular values.
Computing a Reduced-Order Pseudoinverse
In one embodiment, Gr + is computed by taking the pseudoinverse of the transfer function matrix G times the right singular vector(s) corresponding to the “significant” singular values.
Gr +=pinv(G Vr), where Vr is the matrix of right singular vectors corresponding to significant singular values. For the case above, Gr +=[−0.85 −0.59 −0.44]. This reduced pseudoinverse has just one row, so the feedback loop of this example will have just one integrator, the same as the effective rank of G.
In step 505, a transfer function matrix (G) is received, e.g., in a database. In one aspect, this is a commissioning/engineering step that is accomplished by perturbing the actuation of each CRAC unit to determine its effect on the sensors, which is described in more detail below in
In step 510, the singular value decomposition of G is computed, thereby returning U, Σ, and V. U is the matrix of left singular vectors, Σ is the diagonal matrix of singular values, V is the matrix of right singular vectors.
In step 515, a number of significant singular values, Ns, is determined. Ns may be determined via any suitable method.
In one embodiment, singular values that exceed a threshold, such as a pre-specified fraction (e.g., 20%) of the largest singular value, may be selected. An advantage of this method is that it is simple. A disadvantage is that there is no good rule for picking the percent threshold.
Another embodiment chooses the largest singular values of G that result in a Gr + that has row sums (a sum of all of the elements of a row) all with the same sign. These singular values also correspond to the column sums of U all with the same sign. An advantage of this embodiment is that there is a physical interpretation to the threshold. The threshold is the level that (just) moves all actuators in the same direction when all feedback errors are the same.
Yet another embodiment chooses the largest singular values of G that result in a Gr + with all elements the same sign (or all elements of U with the same sign). This method has the same advantages of the second method. It may have a disadvantage of being too aggressive, as it may result in a selection of very few (maybe just one) significant singular values.
Yet another embodiment fits a line to the largest singular values of G as a function of their rank order, and select the largest singular values that do not fit the line. This selection can either be made manually, or it can be automated using an outlier detection method.
In step 520, Ns PID objects are instatiated or otherwise selected for operation. These Ns PID objects will calculate control values corresponding to each of the significant singular values. As such, the Ns PID objects will not be competing against each other.
In step 525, Gr + is computed. In one embodiment, it is not necessary to recompute Gr + from scratch. The singular value decomposition results of G can be used to compute Gr +. Since G Vr=UΣ, then Gr +=Σr −1Ur T, where Σr −1 is the inverse of the diagonal matrix of significant singular values, and Ur T is the transpose of the left singular vectors of G that correspond to the singular values in Σr.
In one embodiment, Σr and Ur, the diagonal matrix of significant singular values and the matrix of left singular vectors corresponding to the significant singular values, are determined, for example by the pseudoinverse calculator 10 or another processor or part of a processor. The inverse of Σr may be taken, e.g., by taking the inverse of each diagonal matrix element (i.e. significant singular value), to obtain Σr −1.
In step 530, control error, Ec, is computed, where Ec=Gr +Ef, and Ef=R−F. R is the reference value for each sensor, and F is the feedback from each sensor. In one embodiment, the control error can be computed by computing Ef first. In another embodiment, r=Gr +R and y=Gr +Y may be computed, and then the difference, r−y, may be taken.
In step 535, the control outputs Oc are computed by the Ns PID objects, i.e. Oc=PID(Ec). These PID objects may calculate the Oc as described herein or known to one skilled in the art.
In step 540, the actuation output Oa are computed from the control outputs Oc. In one aspect, since the system has more actuators than significant singular values (and therefore PID objects), the output of the single PID object is divided among the two actuators.
In one embodiment, the formula Oa=VrOc is used. In other words, the outputs of the PID object is multiplied by the right singular vector(s) corresponding to the significant singular values. If the PID object is a pass-through (proportional gain=1, integral gain=derivative gain=0), then the command to the two actuators of the example above will be
for the above mentioned example. The values in this matrix provides the pseudoinverse (Gc) of the “rank-corrected” transfer function matrix, but by only using with just one integrator, not two, in the feedback loop.
V. Calculating the Transfer Matrix G
It is possible that the actuation outputs, Oa, will be greater than or less than the allowable outputs of the physical devices (e.g., a valve cannot be more than 100% open). In this embodiment, the values of Oa are constrained, and then the values of the integrators of the PID objects are reset so that the integrators don't “wind up”. This constraint mechanism may be combined with any method mentioned herein.
In step 710, the actuation outputs are constrained to Oa,c. In one embodiment, the constraints are maximum and/or minimum values for each actuator.
In an embodiment, a way to constrain the actuation outputs is to fit a line between the previous actuation output vector and the present actuation output vector, and compute the point along the line between them that is at the allowable boundary. The allowable boundary may be any function or shape, and can take account of appropriate values for each actuator.
In another embodiment, another way to constrain the outputs to the final control element is to simply limit the element(s) of Oa to the minimum and maximum allowable values (e.g., 0% and 100%). For example, the value may be for a setting or an actual voltage level. The minimum and maximum allowable values may be chosen based on how the final control element operates. For example, for a valve, they will typically be 0% and 100%, but for a variable frequency drive, the minimum will typically be higher (e.g., 50%) in order to always keep some air moving.
In one embodiment, the constraints are dependent on another variable if the actuation outputs are setpoints. For example, if the supervisory controller 15 resets the return air temperature setpoint of the embedded controller in a computer room air handling unit, the return air temperature setpoint may need to constrained when the chilled water valve is fully closed or fully open, not when the limit of a setpoint range specified by software or by the air handling unit manufacturer is reached. This is because the air handling unit may not be able to achieve extreme setpoints even though them software and the equipment firmware permit them as valid.
In step 720, for embodiments where a reduced-order pseudoinverse (Gr +) is used, the constrained actuation outputs Oa,c are mapped back to the reduced output space oc,c. In one embodiment, Vr is used to map the constrained outputs into the control output space. For example, once the Oa has been constrained to Oa,c, Oa,c may be mapped back to the reduced output space as follows: oc,c=Vr TOa,c, where Oa,c is the vector of constrained actuation outputs. In one aspect, the actuation outputs Oa,c are effectively mapped back onto the output range of the PID objects.
In step 730, the PID objects are reset based on the mapped (if necessary) and constrained outputs. The mapping of Oa into the control output space is used to reset the integrators of the PID objects. In one embodiment, the PID objects are reset so that output values previous to the reset are not used in determining future output values. In other words, the PID object may be reset to start at a first iteration.
V. Calculating the Transfer Matrix G
In step 810, the system is operated in an open-loop state, e.g. each actuator is run at a constant value. In one embodiment, all of the actuators of a same type (e.g. a fan) are run at a same setting (e.g., all blowers at 100% rotational speed). In another embodiment, all of the actuators are run at the same setting.
In step 820, the outputs (e.g. temperatures and humidities) of the sensors 3 are recorded at this quasi-steady state condition. Quasi-steady state refers to the fact that the sensors may not be constantly reading a same value, e.g., due to random fluctuations in the environment or the electrical circuits of the sensors themselves.
In step 830, one of the inputs 12 (actuators) is perturbed by some amount (say, 20%). In this manner, the contribution to G of the actuator whose input is perturbed may be calculated.
In step 840, after quasi-steady state conditions occur, then the outputs are recorded again. Any suitable criteria for determining when a quasi-steady state condition occurs may be used. For example, the changes outputs of the sensors may have a certain rate of change after the perturbation. Once the rate of change decreases below a threshold, then a quasi-steady state condition may be determined. Absolute values for a threshold of changes in the outputs of the sensors is another example.
In step 850, this perturbation is done for each actuator. Each perturbation allows a determination of a column of G. In one aspect, each element of the column being computed is the ratio of the change in the output divided by the change in the input. If the output is temperature, the element of G has the units of temperature/%, when the change to the input is defined as a %.
In step 860, G is calculated from the elements identified for each column of G. The calculation of the transfer matrix G would typically be done once, during commissioning. However, it may be repeated if the system needs to be re-commissioned because of a significant change to the load (e.g. the datacenter).
V. Example System to be Controlled
In one embodiment, fan 950 is a centrifugal fan driven by an A/C induction motor. The induction motor may have a variable speed(frequency) drive VSD 955 for changing its speed. A wireless sensor 960 measures return air temperature, a wireless sensor 970 measures discharge air temperature, and a wireless control module 980 to control the VSD 955. The discharge air temperature sensor 970 and return air temperature sensors 960 may be probes tethered to the wireless control module rather than separate wireless sensors.
In embodiment of operation, the wireless sensors 960,970 send readings over the radio to a wireless network gateway, which passes the signals to a control computer, e.g. which contains supervisory controller 15 of
It should be apparent that various different modifications can be made to embodiments without departing from the scope and spirit of this disclosure. For instance, the invention can be adapted to operate a set of HVAC units that do not have embedded microprocessor-based controllers. In this embodiment, the supervisory controller would provide control commands directly to the final control elements of the HVAC units.
The specific details of the specific aspects of the present invention may be combined in any suitable manner without departing from the spirit and scope of embodiments of the invention. However, other embodiments of the invention may be directed to specific embodiments relating to each individual aspects, or specific combinations of these individual aspects.
It should be understood that the present invention as described above can be implemented in the form of control logic using computer software in a modular or integrated manner. Based on the disclosure and teachings provided herein, a person of ordinary skill in the art will know and appreciate other ways and/or methods to implement the present invention using hardware and a combination of hardware and software
Any of the software components or functions described in this application, may be implemented as software code to be executed by a processor using any suitable computer language such as, for example, Java, C++ or Perl using, for example, conventional or object-oriented techniques. Computer programs incorporating features of the present invention may be encoded on various computer readable media for storage and/or transmission; suitable media include magnetic disk or tape, optical storage media such as compact disk (CD) or DVD (digital versatile disk), flash memory, and the like. The computer readable medium may be any combination of such storage or transmission devices.
Such programs may also be encoded and transmitted using carrier signals adapted for transmission via wired, optical, and/or wireless networks conforming to a variety of protocols, including the Internet. As such, a computer readable medium according to an embodiment of the present invention may be created using a data signal encoded with such programs. Computer readable media encoded with the program code may be packaged with a compatible device or provided separately from other devices (e.g., via Internet download). Any such computer readable medium may reside on or within a single computer program product (e.g. a hard drive or an entire computer system), and may be present on or within different computer program products within a system or network.
The above description of exemplary embodiments of the invention has been presented for the purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise form described, and many modifications and variations are possible in light of the teaching above. The embodiments were chosen and described in order to best explain the principles of the invention and its practical applications to thereby enable others skilled in the art to best utilize the invention in various embodiments and with various modifications as are suited to the particular use contemplated.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US5170935||Nov 27, 1991||Dec 15, 1992||Massachusetts Institute Of Technology||Adaptable control of HVAC systems|
|US5464369||Feb 25, 1994||Nov 7, 1995||Johnson Service Company||Method and apparatus for estimating the rate at which a gas is generated within an enclosed space|
|US5550752||Jul 25, 1994||Aug 27, 1996||Johnson Service Company||Method and apparatus for estimating the rate at which a gas is generated within a plurality of zones|
|US5768121||May 24, 1995||Jun 16, 1998||Johnson Service Company||Adaptive flow controller for use with a flow control system|
|US5862982||Sep 24, 1997||Jan 26, 1999||Johnson Service Company||Optimal ventilation control strategy|
|US5875109||Jul 3, 1997||Feb 23, 1999||Johnson Service Company||Adaptive flow controller for use with a flow control system|
|US5920478 *||Jun 27, 1997||Jul 6, 1999||Oakleaf Engineering, Inc.||Multi-input multi-output generic non-interacting controller|
|US6402043 *||Oct 18, 2001||Jun 11, 2002||John F. Cockerill||Method for controlling HVAC units|
|US6557574||Apr 2, 2001||May 6, 2003||Clifford C. Federspiel||Pressure based flow rate measurement device integrated with blades of a damper|
|US6719625||Sep 23, 2002||Apr 13, 2004||Clifford Conrad Federspiel||Method and apparatus for controlling variable air volume supply fans in heating, ventilating, and air-conditioning systems|
|US6865449 *||May 17, 2002||Mar 8, 2005||Carrier Corporation||Location adjusted HVAC control|
|US7058477 *||Nov 23, 2004||Jun 6, 2006||Howard Rosen||Thermostat system with remote data averaging|
|US7089087 *||May 17, 2002||Aug 8, 2006||Carrier Corporation||Limited access comfort control|
|US7097111 *||Oct 15, 2002||Aug 29, 2006||Gun Valley Temperature Controls Llc||Environmental control system and method for storage buildings|
|US7117129||Mar 11, 2005||Oct 3, 2006||Hewlett-Packard Development Company, L.P.||Commissioning of sensors|
|US7363094 *||Jan 9, 2006||Apr 22, 2008||General Electric Company||Multivariable controller design method for multiple input/outputs systems with multiple input/output constraints|
|US7664573 *||Sep 29, 2004||Feb 16, 2010||Siemens Industry, Inc.||Integrated building environment data system|
|US7839275 *||Nov 9, 2005||Nov 23, 2010||Truveon Corp.||Methods, systems and computer program products for controlling a climate in a building|
|US7847681 *||Mar 24, 2008||Dec 7, 2010||Johnson Controls Technology Company||Building automation systems and methods|
|US20020020446||Apr 2, 2001||Feb 21, 2002||Federspiel Clifford C.||Pressure based flow rate measurement device integrated with blades of a damper|
|US20030064676||Sep 23, 2002||Apr 3, 2003||Federspiel Clifford Conrad||Method and apparatus for controlling variable air volume supply fans in heating, ventilating, and air-conditioning systems|
|US20060116067||Dec 1, 2004||Jun 1, 2006||Federspiel Clifford C||Method and apparatus for determining critical pressure of variable air volume heating, ventilating, and air-conditioning systems|
|US20060206291||Mar 11, 2005||Sep 14, 2006||Bash Cullen E||Commissioning of sensors|
|US20060234621 *||Mar 10, 2006||Oct 19, 2006||Desrochers Eric M||Multipoint air sampling system having common sensors to provide blended air quality parameter information for monitoring and building control|
|US20110161059 *||Jun 30, 2011||Ankur Jain||Method for Constructing a Gray-Box Model of a System Using Subspace System Identification|
|EP0501432B1||Feb 25, 1992||Sep 7, 1994||Hitachi, Ltd.||Method of controlling an air conditioning apparatus and air conditioning apparatus using the method|
|WO2006099337A2||Mar 10, 2006||Sep 21, 2006||Aircuity Inc||Multipoint air sampling system having common sensors to provide blended air quality parameter information for monitoring and building control|
|1||Notification of Transmittal of the International Search Report and the Written Opinion of the International Searching Authority of International Application No. PCT/US2009/035905 dated Sep. 25, 2009.|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US8600560 *||Aug 3, 2012||Dec 3, 2013||Synapsense Corporation||Apparatus and method for controlling computer room air conditioning units (CRACs) in data centers|
|US8924026||Aug 22, 2011||Dec 30, 2014||Vigilent Corporation||Energy-optimal control decisions for systems|
|US20120303164 *||Aug 3, 2012||Nov 29, 2012||Synapsense Corporation||APPARATUS AND METHOD FOR CONTROLLING COMPUTER ROOM AIR CONDITIONING UNITS (CRACs) IN DATA CENTERS|
|U.S. Classification||700/276, 700/29, 340/514|
|Cooperative Classification||G05D23/1931, F24F11/006, H05K7/20836, G05D22/02|
|European Classification||F24F11/00R5, G05D23/19G4B, H05K7/20S90, G05D22/02|
|Apr 9, 2009||AS||Assignment|
Owner name: FEDERSPIEL CORPORATION, CALIFORNIA
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:FEDERSPIEL, CLIFFORD C.;REEL/FRAME:022525/0519
Effective date: 20090303
|Jun 17, 2011||AS||Assignment|
Effective date: 20110426
Owner name: VIGILENT CORPORATION, CALIFORNIA
Free format text: CHANGE OF NAME;ASSIGNOR:FEDERSPIEL CORPORATION;REEL/FRAME:026460/0555